The deformation of Australia observed with InSAR

John Dawson 1 , Paul Tregoning 1

1 Research School of Earth Sciences, Australian National University , Canberra , ACT 0200, Australia

We assessed the accuracy of earthquake source parameters inverted from simulated Interferometric Synthetic Aperture Radar (InSAR) data. Using focal mechanisms of Australian earthquakes (1959 to the present), we simulated synthetic two-pass InSAR observations with realistic spatial noise derived from the characteristics of actual ERS-2 and ENVISAT InSAR data observed over Australia. The precision of two-pass satellite SAR interferometry with ERS2 and ENVISAT SAR data in the Australian region can approach ±2 mm (1 sigma) and is routinely at the ±4 mm level. The use of spatially uncorrelated observational weights has minimal impact on the accuracy of earthquake source parameters inverted from InSAR data but invalid a priori assumptions of the dimension of the earthquake rupture plane can bias depth estimates by up to 0.4 km. In most cases single geometry (i.e. ascending or descending) InSAR observations can be used to accurately determine earthquake source parameters, although typically two-pass geometry reduces the source parameter uncertainties by a factor of 1.5. In general, earthquakes of magnitude < 4.8 are unlikely to be detected by InSAR although deformation from very shallow events would be visible. InSAR is insensitive to magnitude 6.2 earthquakes deeper than 10 km, and magnitude 5.5 deeper than 6 km. For earthquakes magnitude > 5.8 (average depth 6.5 km) we could estimate the epicentre of the rupture with an average accuracy of 0.25 km, depth to within 0.5 km and the fault orientation to better than 2 degrees.

Our ongoing research will focus on temporal image stacking and techniques for the identification of Persistent Scatterers (i.e. temporally phase stable radar reflectors) applied to the observation of ‘slow' (~1mm/yr) deformation processes. Our research will focus on new strategies for the identification of persistent scatterers and the application of new modelling techniques to regions where InSAR is known to not work as effectively, including humid and heavily vegetated regions of Australia and Papua New Guinea . The principal motivation and contribution of our research is to: a) demonstrate, 'tune' and develop new PSInSAR techniques, for the observation of 'slow' geophysical deformation phenomena in InSAR-adverse regions; and b) improve the accuracy, precision and computational efficiency of these techniques.

Figure 1. Simulated inteferogram of surface deformation of the 1968 Meckering earthquake. Line-Of-Sight interferometric observations with spatially correlated noise from descending two-pass satellite InSAR (ENVISAT). Each colour fringe cycle represents 28mm of deformation. The centre of the earthquake rupture plane is located in the centre of each image.